WEEKLYThis week's high-risk CVEs and how to respond | RAIMEI weekly report
ProductsRAIMONAI-WAFcmdAI PentestHONMARUAI Ops PanelSHIELDMobile App ProtectionKABUTOPC Software ProtectionSENRIGANAttack Surface Management (ASM)TAIKOUptime MonitoringINAZUMAPenetration TestingRAIMEIVulnerability IntelMOBILEMobile Security App
SolutionsGamingFinance & paymentsE-commerce & retailSaaS & startups
DevelopersDocsAPI referenceCLI・SDKCI/CD integration
PricingVulnerability IntelFree pentestSupport & contact

cmdcmd.jpEarly access

cmd AI Penetration Testing
Automated penetration tests run by an AI agent

cmd is an AI agent that runs penetration tests on its own. It actually attempts attacks and finds the weaknesses that can really be exploited.

  • Command line and web console
  • Billed by token usage
  • Only tests what you own
cmd — pentestSample
$ cmd pentest --target https://staging.example.com > verifying ownership ... ok > recon: 14 endpoints, 3 forms, 1 login [done] authentication checks [done] injection checks [done] access control checks > 2 findings · report ./cmd-report.html
Runs after every release

It can be triggered from CI/CD, with no expert booking required.

Reports with reproduction steps

Each finding lists its risk level, reproduction steps and a suggested fix.

Only tests what you own

Before a test, cmd verifies that the domain is yours. If verification fails, the test does not start.

Billed by token usage

You pay for the tokens a test actually consumes.

HOW IT WORKS

How a test run proceeds

Four stages, from ownership verification to the report.

  1. Verify ownershipcmd confirms the target domain belongs to the requester before testing starts.
  2. ReconIt enumerates endpoints, forms and login entry points to define the test scope.
  3. Attempt attacksIt checks authentication, injection and access control one by one and confirms whether each weakness is exploitable.
  4. Produce the reportIt outputs an HTML report listing findings, reproduction steps and suggested fixes.

This flow is an architectural sketch. See the official documentation for details.

COMPARE

How the three testing methods differ

cmd suits routine verification after each release. For complex business-logic flaws, we still recommend verification by INAZUMA experts.

Vulnerability scanINAZUMA expert pentestcmd AI pentest
Run byA scanning toolSecurity expertsAI Agent
MethodMatches known vulnerability signaturesTests with an attacker's methodsMaps the target, then attempts attacks to confirm exploitability
When it runsOn a scheduleBy appointment, scheduled by scopeAfter each release, triggered from CI/CD
Best forFinding known weaknesses broadlyIn-depth verification before launch or after major changesRoutine verification after each release
Billing—Quoted by scope; the first test is freeBy token usage

FEATURES

What cmd checks and provides

Feature list includes planned items

Automated recon

Enumerates the target's endpoints, forms and login entry points.

Authentication and authorization checks

Tests whether login mechanisms and access controls can be bypassed.

Injection checks

Tests for SQL injection, command injection and related flaws.

Trigger from CI/CD

Runs from the command line as a step in the release pipeline.

HTML report

Lists findings, risk levels, reproduction steps and suggested fixes.

Asset ownership verification

Confirms before each run that the target belongs to the requester.

USAGE

Two ways to use cmd

Command line

Run from a terminal or CI/CD

One command sets the target and starts the test; the report is written as an HTML file. Suited to release pipelines.

$ cmd pentest --target https://staging.example.com Sample

Web console

Sign in at cmd.jp

Add targets, start tests and read reports in the browser.

Go to cmd.jp

PRICING

Billed by token usage

cmd is billed by the number of tokens consumed during a test. A wider scope and more checks mean higher usage. Unit price not yet published

Request early access

FAQ

cmd FAQ

A vulnerability scan matches known vulnerability signatures. cmd actually attempts the attack, confirms whether the weakness is exploitable, and includes reproduction steps in the report.

Not entirely. cmd suits routine verification after each release. Complex business-logic flaws and items that need deeper verification can be handed to INAZUMA experts.

No. Ownership must be verified before a run, and targets that fail verification are not tested.

We recommend running against staging first, then deciding whether to run against production once you have reviewed the results.

By the number of tokens a test consumes. The unit price has not been published yet; during early access, please ask by email.

cmd, let AI be your red team

cmd is in early access. Apply by email and we will reply with the scope and timing we can offer.